calibration of measurement errors

[bmicallef]

I have estimated two-country SOE in a currency union a la Gali/Monacelli with habit formation and price indexation. However, since the model is highly stylized, measurement errors tend to be large. Is there a way in Dynare to calibrate the variance of the measurement errors to reasonable values, so that, the fundamental shocks explain, say, 70%-80% of the variation in the data.

Thanks
B.

[jpfeifer]

Usually you use priors for that. Use e.g. a uniform prior for the stderr of the measurement error and set the upper bound to the desired value, i.e. fraction of the data variance (which you have to compute yourself).

[SébastienVillemot]

You cannot do that automatically.

What you can do is to estimate the measurement error using a tight prior distribution (this is almost equivalent to calibrating), see if the results are in line with what you want and, if not, refine and retry.